Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{208612,
        author = {Akanksha Patil and Sujata Sathe},
        title = {Blockchain-Based Data Provenance and Integrity Verification for Secure AI Training},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {no},
        pages = {603-609},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=208612},
        abstract = {The growing use of Artificial Intelligence (AI) has made the quality and reliability of training data increasingly important. If training data is modified without authorization, contains malicious records, or includes incorrect information, it can affect the performance and reliability of an AI system. 
These issues can lead to data poisoning, where compromised data influences the training process and produces unreliable results. Existing centralized methods for tracking data provenance may not provide sufficient transparency and can be vulnerable to tampering or a single point of failure. This paper proposes a blockchain-assisted framework for tracking and verifying the integrity of data used in AI training pipelines. The framework records important information such as data hashes, source details, timestamps, dataset versions, and modification history on a blockchain. This creates a transparent and tamper-evident record that can be used to verify whether data has been changed. Smart contracts are used to enforce predefined rules for registering and modifying datasets. 
Before data is used for AI training, its integrity is checked against the information recorded on the blockchain. Any unauthorized change, hash mismatch, or suspicious data submission is flagged for further verification. The proposed framework is evaluated using controlled data-manipulation scenarios, with performance measured through poisoning detection rate, verification time, provenance traceability, transaction latency, and storage overhead. The study explores how blockchain can provide an additional layer of trust and transparency to AI data pipelines while making compromised data easier to identify and trace. The implementation is available at https://github.com/akankshapatil99/BAAITF.},
        keywords = {Blockchain technology, artificial intelligence, security, integrity, data poisoning},
        month = {September},
        }